#005: Big Data -- What an Executive Needs to Know

The power of big data is a curious thing, Make a one man weep, make another man sing. Change a hawk to a little white dove. More than a feeling that’s the power of big data. As always, Huey Lewis hits the nail on the head with this complex topic. What does the phrase actually mean? How can my company take advantage of it? Michael, Tim and Jim take on big data in episode five, and try to focus in on making this hard to pin down concept understandable and relevant. All this and more in one American hour, 46 Canadian minutes.

Episode Transcript

The following is a straight-up machine translation. It has not been human-reviewed or human-corrected. However, we did replace the original transcription, produced in 2017, with an updated one produced using OpenAI’s WhisperX in 2025, which, trust us, is much, much better than the original. Still, we apologize on behalf of the machines for any text that winds up being incorrect, nonsensical, or offensive. We have asked the machine to do better, but it simply responds with, “I’m sorry, Dave. I’m afraid I can’t do that.”

00:00:04.00 [Announcer]: Welcome to the Digital Analytics Power Hour. Three analytics pros and the occasional guest discussing digital analytics issues of the day. Find them on Facebook at facebook.com forward slash analytics hour. And now, the Digital Analytics Power Hour.

00:00:25.58 [Michael Helbling]: Hello, welcome to the Digital Analytics Power Hour. This is Episode 5. You know, you’re an executive, or maybe you work for an executive. But one thing your executive, or you, have heard about is big data. That’s today’s topic. Let’s dig in to big data. It usually comes with a big price and a big promise. And we’re here today to talk about some of that. Obviously, I’m going to be joined today by my two other hosts, Tim Wilson, partner at WebAnalytics Demystified. Hey, Michael. and Jim Cain, world traveler and double CEO of Babbage Systems and Napkyn Inc. Hello, gents. And obviously, I’m Michael Helbling, the analytics practice leader at Search Discovery in Atlanta, Georgia. All right, guys. Never before have such great guys undertaken a task for which they were so not prepared. No. In the realm of big data, we definitely have a lot of noise out there and when you really start to test it, sometimes it seems like the promises are a little too good to be true and the excitement around it is a little more hype than, I can’t think of an H word, but basically substance. Right? So guys, maybe to start the show, let’s dig into the topic of or the definition of big data and start there. But so let me hand it off to you guys.

00:01:59.52 [Tim Wilson]: Who wants to take it first?

00:02:00.48 [Michael Helbling]: Yeah, come on.

00:02:04.25 [Tim Wilson]: I’m perfect to go second. I mean, so to me, big data, it is a term that anybody and everybody has free license to use however they want. And so I kind of get struck when I hear a And I think there’s even kind of a standard kind of normal way for people to talk about big data saying big data is not just more data. But a lot of times that’s what it gets treated as is big data is more data. So I think there are probably two or three or there’s a big data number of competing definitions for what big data is. But one person who kind of talked about it and struck me as being kind of unique but I think he was kind of being contrarian was you know, Gary Angel from Symphonic or Ernst & Young. And he, and I may completely butcher his definition, he kind of goes through the four V’s of data and says, look, it’s not because it’s just more data. It’s not because it’s real time. He said, you know, to him, the big distinguishing characteristic is that it’s sort of the, it’s the combination of the sequence of things happening. and the volume of what we actually know about what’s happening, which made me think a little bit. But yet when I go to Wikipedia or when we go to Webster’s or whatever, I don’t know that that definition is necessarily there. So I think, Jim, maybe now it’s your turn to define big data.

00:03:33.56 [Jim Cain]: Well, it’s just interesting that three guys that love to have an opinion all dodged on being the first to say what it means, you know?

00:03:39.45 [Michael Helbling]: Because, well, everyone knows that whoever opens their mouth first is going to be the primary target of all the backlash.

00:03:46.31 [Jim Cain]: Yeah, you’re the column A vendor for all the trashing. But, I mean, the way that I describe Big Data to people is just all eggs, one basket. I mean, I know it’s kind of idiot sales guy addition, but if you take all the data in your business and all the data that you’d like to capture and you put it all in one spot, then you can reap the benefit of the network effect of all of those things in one place. And I don’t think it’s really that much more complicated than that until you start getting to very specific use cases and verticals and types of data. All your eggs in one basket.

00:04:17.68 [Tim Wilson]: Am I nuts? Well, but it seems like there’s a lot of the definition is around the technology that you’ve got when Hadoop comes out when there’s some level of definition that big data is when you’ve got more data. So if you put it all in one basket, but then you don’t have the horsepower to actually crunch it because you can’t just throw it into big. you know, flat tables. So you put it into Hadoop, into some new sort of data warehouse that they can crunch it. And it kind of sticks in my craw when it gets defined as something that requires more technology or a different approach to crunch it. So I do feel like there’s a legitimate definition that’s more than that. But I think that’s where a lot of definitions are.

00:05:02.58 [Michael Helbling]: We did Google it and Google says it’s extremely large data sets that may be analyzed computationally to real patterns, trends, and associations, especially relating to human behaviors and interactions. We know it’s interesting.

00:05:18.74 [Tim Wilson]: You start digging into some of the use cases or where the definitions of it. So you hear human behavior and interactions, and you think, oh, this is all about marketing. And you read elsewhere, and a lot of the talk about big data is, No, it’s like crazy hyper volumes used for science and engineering and pattern matching of things that aren’t related to human behavior. I look at the human behavior part and humans are so messy and so noisy and so not repeatable. It’s so much easier to predict what a paramecium is going to do than it is to what a human is going to do.

00:05:58.52 [Michael Helbling]: Yeah, I mean, and big data, you know, could be the migratory patterns of birds, you know, correlated with the ice cream consumption in Ottawa, you know, I mean.

00:06:09.47 [Jim Cain]: Well, there’s none right now.

00:06:11.14 [Michael Helbling]: OK, well, there you go. But yeah, I mean, so this is going to be a pretty lame show if we can’t move past the definition, though.

00:06:20.45 [Jim Cain]: Can we agree with my statement that all the data needs to be in one spot? Can you use our Hadoop to go to seven different disparate data sources and do stuff? It’s all going to be one spot for it to be big data, right?

00:06:35.24 [Michael Helbling]: I don’t know. I think so. I feel like that is a description of data warehousing, which could definitely be part of big data. That’s the thing is, I think maybe big data tries to not end with what does your data store look like? It tries to keep going into, and we’re going to apply these sorts of computational analysis to it, and we’re going to apply these kinds of algorithms to it. And big data becomes something well beyond the data itself. It’s the data itself and the functions behind it. And maybe that’s really where the problem lies, is people are describing multiple fields in terms of where data gets stored and how it’s stored and how much of it there is, and the methods used to parse and function within that data to analyze it.

00:07:26.77 [Tim Wilson]: I wonder if some of it is the evolution that if you look at how we got to this concept of big data was this explosion, and it wasn’t marketing and digital and the interactions we can capture, but I think it’s also in instrumentation and the ability to record temperature changes at the nanosecond, for all I know, and the volumes of data that generated. And there was kind of a march to the data’s to not granular enough. Technology came along to capture more granular data. That was seen as a technology problem. Then it was, well, if we can capture it, we’ve got it stored somewhere. So that became a storage problem. And then it became a, that data is arriving so fast. It’s not just storing it. It’s actually being able to write it fast enough. And I think that was kind of one generation, that that got solved. That it was, we can now write this stuff fast enough we can capture it. And there was an assumption that that was a good thing. And then all of a sudden, people said, oh, I want to write a SQL query against this. Holy crap, I have two million rows in this table that I’m trying to join to two million rows in this other table and I can click go and my query is gonna run for a month and a half and it might not return the right stuff. And that’s when all of a sudden this, it emerged, and this is me speculating, I am not a history of the etymology and of big data, was all of a sudden there was this aha, that oh crap, the challenge was the biggest hurdle wasn’t actually capturing or storing the data, it’s now how do we actually manipulate that data in a way that is reasonably timely. You know, I can do a VLOOKUP on 10,000 rows of data in Excel, and it’s going to return effectively instantaneously. And when Excel was in Excel 2007, I think, that went up to a million rows, and it still worked pretty well. But not if I was trying to do VLOOKUPs from a million rows to a million rows. And all of a sudden that was kind of, so it sort of came from this technological basis that I know in theory I can join this stuff. And so I do tend to agree with Jim that I think there is an implicit assumption that data has to be kind of aggregated somewhere that is accessible because otherwise there’s no way that you’re going to be able to join it. And maybe that’s kind of a key to the definition is that it’s not one data set, it’s joining multiple large datasets and that’s where people perceive the value coming from.

00:10:03.50 [Jim Cain]: So is the fact that we’ve been talking for 10 minutes and none of us can pin anything down other than, yeah, it’s probably all in one spot. Is big data a synergy word? Is it a bridge the gap? Is it a guru evangelist word? Does it mean anything?

00:10:18.61 [Michael Helbling]: Well, and I think that’s kind of what we want to discuss, right, is how do we kind of clear away some of the brush around the actual truth of it, right? which is, well, okay, so now that you sort of have a working definition maybe of it’s too big to handle. I’m making one. Give me a second. It’s like, you know, data too big to be analyzed in classical SQL ways and needs to be held in one kind of data store. You know, that could be our definition for today of big data. I’m sure there’s about five people I’m thinking of right now who are gonna believe that’s a woefully inaccurate description. However, if we work with that just by itself, then we can go back to and say, okay, now, how do we bring any value to an organization with this? What does it mean to use big data?

00:11:09.64 [Tim Wilson]: So I think that is actually, with that definition, even if it’s wildly imperfect, what that sets up is that Yes, there are situations where that scenario exists and there is a clear sort of path to value to be gained from it. But that scenario does not apply to 95% of the scenarios where people think it should apply. When we’re looking at Web Analytics data, it’s not big data. If we’re a massive website and we’re getting The raw log files, yes, it’s big. eBay has big data. Amazon has big data. A $5 million a year e-commerce company, I don’t think they do.

00:11:55.81 [Michael Helbling]: That’s interesting that you say that, Tim, because you initially brought up Gary Angel and his definition of big data, I would say, actually does apply to just analytics data. Because he’s talking about it from the concept of the stream of interactions as opposed to a discrete set of aggregated transactions, which is more of a classical BI.

00:12:16.36 [Tim Wilson]: But I think about it, but I think you have to have you have to have sufficient volume that you can actually identify patterns in it and we’re talking about people and their behavior and it goes back to 10 years ago when people thought that a clickstream analysis, we’re going to find the most common paths through our site. And you’re like, yeah, the most common path happens 0.1% of the time. So that’s good. Just a fair criticism.

00:12:39.57 [Jim Cain]: The session that we had, I don’t know, is one or two episodes ago when Michael was like, what do you think about the concept of capture everything now and figure it out later? And you and I were like, hell no, hell damn no. That’s sloppy analysis. I mean, that concept is pretty much what we’ve already been talking about. Like that’s what we’ve gotten to in terms of our big data definition. So now I’m wondering if we say, you know, big data is the ability to provide business value or to answer a business question. Well, I think through the application of advanced data mining technology on a significant volume of data that’s in one place. Like we’re getting there, but I think big data equals business value.

00:13:22.33 [Michael Helbling]: But well, hold on a second. I think big data is just big data. The business value comes from how you leverage big data, because there’s plenty of people who are going to do big data and already are. but aren’t getting Jack squat in terms of value out of it. It’s not because they’re not trying to do big data or they’re not using big data or big data stores or whatever you want to call it. But that’s the, I mean, and this is kind of where it meets kind of what I think of as, okay, if we’re talking to a C level executive right now, we’ve got 30 seconds to help him understand what he needs to know about big data and how he’s going to apply it to his business. How would we answer that question?

00:14:00.74 [Tim Wilson]: So I think that’s simple. is you don’t spend a dime until you have articulated a list of problems that it’s going to solve. And I think this was, Michael, you’d found this beforehand, the red herring of big data, the post, and I’m trying to scan it now where he was saying this is this perception that, oh, you don’t have to come up with hypotheses. You don’t have to ask questions. You just take this massive volume of data And it’s going to find stuff for you. And that has bugged me. I can point back easily 12 or 13 years to having a debate at a consortium, a little small thing at Texas A&M, with some other people who said, no, no, no, if you just point these things at your large data set and they find correlations, then you don’t have to come up with hypotheses. And that’s a total load of crap that if I was talking to an executive, I’d say, what question can you not answer? And now let me take that question and say, would big data reasonably answer that? And how much am I going to have to invest to get that data and join that data? And I think when you take the target pregnant teen example, They had a very specific thing they were trying to do where they said, we have enough evidence that these buying patterns are indicative of characteristics of our consumers. And then they went out and invested in what they were doing. And Nordstrom’s kind of the same way. They started with their strategy of saying, we’re going to empower for time immemorial Nordstrom’s has empowered their employees to have phenomenal customer service and so they came at it and said, how can more data further that specific objective? And I would guess that they actually sat down and had that documented and clearly defined. So they had some direction. They didn’t just say, hey, go out and capture all the data and then we’ll figure out what to do with it.

00:16:06.06 [Michael Helbling]: And this gets to the crux of where the problem is with big data, I think, which is that this you know, magical computer savior is going to come along and fix all of our business woes, and it’s a fence to science a little bit. It’s the kind of religious fervor that goes into this kind of thing, you know?

00:16:27.03 [Jim Cain]: Well, I for one welcome our new robot overlords. That’s good.

00:16:34.49 [Tim Wilson]: You’re sucking up the digital technology we’re using to record this podcast. And hello to the NSA while we’re at it.

00:16:40.40 [Jim Cain]: Oh, yeah. There you go. Now, I want a T1000. That’d be so cool. As you were talking through that, your example, Michael, and the thing about Nordstroms, now I wonder, I just keep trying to refine. Because we keep saying, I don’t know what the shit big data is. And then if I was talking to an executive, I’d say, here’s when you can use big data, but we still can’t define it. Is big data really just stuff a data scientist does?

00:17:06.19 [Michael Helbling]: Well, then we’d have to define data scientists.

00:17:08.45 [Jim Cain]: That’s what I can do.

00:17:09.71 [Michael Helbling]: That’s a whole other podcast.

00:17:11.44 [Jim Cain]: It is, but I can define data scientists, but I can’t do big data.

00:17:15.34 [Tim Wilson]: Is the data scientist a growth hacker? Because we could just kind of keep heading down this path.

00:17:21.03 [Jim Cain]: You give it away episode six, which will be all the stupid titles we’ve ever heard.

00:17:28.60 [Tim Wilson]: I agree that you can define big data and Michael, your definition, I think is probably as workable and non-controversial as you’ll get of one around large data sets being joined together. I think it was a fair take that it doesn’t, unfortunately, there is not part of the definition that it has to deliver business value, but that’s what we would want to say to executives, that you’ve got to figure out what that business value is or has a high likelihood of being before you actually,

00:18:01.31 [Michael Helbling]: with big data consider for a moment that that executive is already inundated with reasons why if he’s not on the big data train he’s been left behind right his business will suffer because he’s not competing with big data or whatever the terminology of the day is so then if you’re talking to that guy he’s looking for Okay, everybody is telling me that if I’m not doing big data, whatever that means, and the definition is pretty amorphous as we’ve proven tonight, I’m getting left behind. I’m not getting a competitive advantage anymore. I’m being left behind by my competitors.

00:18:38.98 [Jim Cain]: Yeah, but that’s old school fear, uncertainty, and doubt selling. That’s how large software companies, and I’m not going to name any names IBM, sell significant contracts to companies.

00:18:51.94 [Tim Wilson]: I didn’t name any names. But if the executive comes to me and they say, look, tell me what I need, and I say, if my response is, we need a one day off site and we need that one day off site exactly two weeks from today and you need to tell everyone who’s going to be at that off site that they need to come with a list of large data sets they have and ideas for what those data sets could do, that to me would be a reasonable, hey, if you’re scared, then what I need is a day of your focused attention and I need to sit down with the whiteboard and we need to draw out what can we actually have? Are you in financial services? What do we have in the way of credit score data and web behavior data and our own customer record data and geo data that we can stitch together. And let’s draw those as big boxes that say, I’m going to join massive, dirty, complex data set A to massive, complex data set B. And it’s a noisy join. It’s probabilistic. It’s not a perfect key. And if I do that, here is something I can do to differentiate.

00:20:03.66 [Michael Helbling]: And how much would you charge for a day like that, Tim?

00:20:09.93 [Tim Wilson]: If the inner voice wasn’t saying, oh, Jesus Christ, I sound like a consultant again.

00:20:15.34 [Michael Helbling]: I don’t care who does that. Honestly, I think there’s probably some problems.

00:20:22.69 [Tim Wilson]: I would pay for that. I would pay for that. I would pay. I would pay four figures to go and have that conversation with an executive who’s not a client because I would actually get to have that moment of saying, no, you actually need to sit down and be really clear on what your strategy is and what your competitive advantage is or could be. And I will happily do that. I would love to do that.

00:20:44.47 [Michael Helbling]: Well, and I think the three of us being fairly pragmatic fellows, typically, I think this is what befuddles us the most, if I can speak for all of us, about where we see big data in the enterprise and how it’s being handled. Because the pre-thought around how do I actually get something valuable out of what is actually a pretty big investment seems to be either merely lip service or almost no planning or thought at all. Maybe that’s the issue we’re really trying to grapple with and if there is an executive listening, that’s our encouragement to you is, you know, do that upfront thinking.

00:21:24.13 [Jim Cain]: I think we’re skirting around the statement that big data in the absence of a really high opportunity cost business case is kind of bullshit. Big data is not a thing that you buy for later. For the same reason that collecting all your data now because maybe you’ll need it later is a bad idea and I can give several examples of the last couple of years of being on the phone with a senior decision maker who just made or was about to make a very expensive consulting or software decision because they felt they need to keep up with the Joneses. And after having a conversation about the time, the complexity, the cost, and the lack of defined business value, we ended up in a conversation. You guys have had this conversation before. Remember, have you ever seen the Louis C.K. routine where his daughter keeps saying, why? She says, why Skype movies?

00:22:17.21 [Jim Cain]: Yeah, yeah, why?

00:22:19.17 [Jim Cain]: Why? Why? I don’t remember why, because I smoked a lot of pot in high school. Why? Because I was really unhappy. He keeps saying why to a senior executive about big data. We’re going to spend a ton of money on XYZ company? Well, why? Well, because we need to get into big data. Well, why? Because we’ve got data all over the business. And you get to a point where they go, I’m not sure. But personally, I have a specific thing I would like to learn. And three out of three times where I’ve had that conversation, there was a much easier way to solve that problem. And there wasn’t a use case at that time for a million dollars to build a super hypothesis machine.

00:22:59.96 [Tim Wilson]: So it’s interesting, you’re fed with, I’m going to jump into it because I’ve got to start, and we had this discussion when we were talking about collect everything, it was, I’ve got to start collecting it now because whenever I figure out how to actually crunch it, I’m going to need some large amount of history to work with that. And I think that is actually a red herring as well because in the world we live in, the fact is, I’ve never found use in web analytics data that’s three years old. If I’m trying to crunch big data of weather patterns, absolutely. I want to capture more detail and as many data points as possible so that I have models I can feed into because although weather patterns may be shifting, you know, historical, I can see capturing that data. But for our purposes, once you get past almost it’s 13 months, right? It’s once you’ve gone Who looks back at your comps over two holiday seasons ago? It doesn’t happen. So I’ve got to just focus on capturing the data because I’m going to need as much of it as I possibly can get is like, no, you’re not going to need to wait that long once you started really gathering that larger data set before you’ve got enough for you to kind of mine and work with.

00:24:14.25 [Michael Helbling]: So I think there’s a lot of different things that could be done with big data. And I want to turn our conversation to those. What are the valuable things that businesses can use big data for?

00:24:25.28 [Tim Wilson]: Well, so I’m assuming we need to kind of scope it to the… Well, this is the digital analytics power hour.

00:24:33.73 [Michael Helbling]: So let’s try to stay close to digital data.

00:24:38.13 [Jim Cain]: Well, I mean, I can give you an example. In one scenario, we had a customer in apparel. And the consumer profile for people who bought in store, because they had a large network of stores, a lot of stores and a catalog and a website. And they spent a lot of money on offline profiling for demographics. People who buy in store, totally different demographic profile than the people who buy in the catalog. And the people who buy online are totally different as well. So three totally different groups of women interacting with their business. They’re sitting back saying how do we streamline purchasing and marketing so we don’t need to run three separate businesses? To me that’s an interesting that that is a big It’s a it’s a very good question with some key defined outputs And it’s the kind of scenario where you sit back and you probably hit a wall where you say we need to start thinking about advanced data capture data analysis tools big data

00:25:39.28 [Tim Wilson]: So now’s what I want to call Kevin Hilstrom and just get him to give the answer to that question in the absence of big data. I’m not sure that’s a… I mean, it does, as you were talking, I’m thinking that is the sort of stuff that Kevin talks about, that people talk about omni-channel like it’s one consumer freely floating across all these channels and in a lot of ways it tends to not be the the case, but if you sit back and say, what could we do? Do we understand who these are? Do I need big data to go out and survey a thousand of each of those and actually understand their attitudes and behaviors and then do some analysis on it and say, these are fundamentally different and I can or I can’t merge them? To me, that wouldn’t necessarily

00:26:28.76 [Jim Cain]: qualify as a… No, the surveying is pretty straightforward, but then the surveying fed the hypothesis and the hypothesis was that if we streamline all of the products we carry and the way we talk about them, we should be able to create a more streamlined experience that makes the most money possible with three totally different groups of women. Yeah. That’s the big data.

00:26:55.06 [Michael Helbling]: I think at the end of the day, like big data, well, so I’m going to start with a massive oversimplification, which is web analytics or digital analytics is mostly about trying to find the difference between two things, right? And then big data is that same pursuit, but also trying to predict the future based on the understanding of that difference. Whether that be the type of customer, the type of transactions, the type of interactions, the behaviors, the attributes, whatever it is about the things that we’re analyzing.

00:27:34.25 [Tim Wilson]: There’s the potential application of big data for predictive analytics. Can I spit out what levers I can pull and predict how that is going to change business results? I actually think there’s an operational component that, and this harkens back to some stuff that we tried to do 10 years ago, very, very ham-handedly without the rapid processing. But if I’ve got volume attraction in my site, and I have IP address, and they’re not logged in, but I can use 25 different variables to try to kind of hone in on who do I think this person really is. And then if I’ve got in-store data, And that in-store data is collecting some level of transactional, behavioral app usage. I don’t know what. If I can take these two data sets, and then I can take third-party, experienced data, and kind of stitch that onto the households, and if I can combine all of that and say, I think I can get a critical mass of being able to understand with reasonable certainty who I can make offers to. So take the target. the quick for anyone listening to this who hasn’t heard the Target example. So Target, that example of where they were printing their flyers and sending them and they were personalizing the flyers, they weren’t personalizing them, they were sending different subsets of their flyers to different stores. They had figured out that based on buying patterns, they thought they could detect when someone was pregnant, a guy goes in, complains to his local Target and says, why are you sending these flyers to our house addressed to my daughter that our pregnancy stuff, my daughter’s not pregnant. The store manager apologized, felt terrible. The store manager, I guess, followed up a couple weeks later and said, just following up again, wanted to apologize for what happened. And the father said, apparently there were some things going on in our house that I didn’t know about. And it turns out my daughter is pregnant. So the upshot was that Target’s data mining figured out that this guy’s daughter was pregnant before he knew it himself. That is a reasonable use case that we have a finite set of activities that our consumers are doing with us and we think that we can predict future behavior and if we can predict the future behavior or future characteristics of that person then we can make offers to them that they are more likely to respond to. But that’s operational. You have to build the model and say, yeah, this actually holds up. It holds up enough. But then ultimately, the challenge is crunching that data fast enough that you can get the customized flyers printed or the customized emails made and out the door. And they’re still not going to be perfect. And those are going to work better than just doing things with kind of spray and pray type marketing.

00:30:28.99 [Michael Helbling]: I agree. But I don’t see how that breaks down my oversimplification at all. because we’re just trying to figure out the difference between pregnant and not pregnant and then do something different or understand something different about the consumer at that point, right?

00:30:45.40 [Jim Cain]: Well, and I think the thing that Tim did was he also supported your statement about effective big data is something that’s future-looking. Like, analysis is ghost of Christmas present and ghost of Christmas past, and big data is ghost of Christmas future.

00:31:01.58 [Michael Helbling]: But I mean, in a certain sense, all of that we do in digital analytics is about predicting the future. We look to the past to figure out a path to the future. I mean, this is why my studies in history have served me so well. I’m a data historian, right? I look at that and look at the patterns of the past to understand what will happen in the future.

00:31:25.77 [Jim Cain]: I would like your LinkedIn title to change immediately.

00:31:30.82 [Tim Wilson]: Well, but I would stop short of that. I mean, we like to say predictive. I’ll fly out and say I don’t necessarily build predictive models. But if I can look in the past and say this absolutely did not work, so let’s not do that again. That’s kind of different from saying that’s a prediction. Well, but I’m not quantifying. You’re not quantifying the result.

00:31:59.07 [Michael Helbling]: So maybe that’s the difference between like, hey, don’t be a dummy and do the same thing as broken versus this will go up by 30% if you do X, Y, and Z. Yeah, I’m willing to go there with you. That’s good.

00:32:13.17 [Jim Cain]: And I think a well-executed big data practice can accurately predict in a way that’s not just directional, specific things that are going to happen that will make you more money. So I’m entirely disagreeing with your statement, Michael. It’s not only future.

00:32:27.76 [Michael Helbling]: But I think, you see, this is where I feel like, you know, the minority report thing starts to happen where we’re like, yeah, the future is all of us sort of like envisioning little things and we’re slicing and dicing data with our fingertips in the sky. And, you know, we’re right back into the computational savior mode, right? Or on our way there.

00:32:51.10 [Jim Cain]: So I don’t know. That’s my new LinkedIn profile.

00:32:53.68 [Michael Helbling]: Computational savior.

00:32:54.96 [Jim Cain]: Heck yes.

00:32:55.62 [Michael Helbling]: I like it. Here’s the thing I heard that I think I want to reiterate and then I want to riff on for just a second. And that is figuring out and defining the value of big data before you jump in with both feet. I’m pretty sure I heard us say that tonight.

00:33:13.23 [Tim Wilson]: Cannot disagree with that. I think that is

00:33:16.86 [Michael Helbling]: Tim will sit down with your organization for a day and help you do that value add for free. You heard it here.

00:33:26.57 [Tim Wilson]: Let’s say if you’re a hundred million dollar or more organization, I will pay fifteen hundred dollars to sit down and do that with your CEO and CMO. You’re crazy.

00:33:41.94 [Michael Helbling]: If anyone is listening and has the ability to put those two people in a room and you don’t do that with Tim Wilson, you’re a moron.

00:33:50.91 [Tim Wilson]: I will not pants stress and I will also spend a ton of time prepping for that, but I would absolutely love the opportunity.

00:33:58.60 [Michael Helbling]: Buy now and leave. That’s right, but that’s not all. There’s more.

00:34:03.41 [Jim Cain]: I like doing the kind of periodic recaps, but we started off with no one wanting to, like we threw Tim under the bus with, I don’t know, Tim, what do you think big date is? And then I think we’ve gotten to a place where we agree that it’s a massive data set that lives in probably one place. And we’ve also agreed that to justify big data, big data requires a business case. Is there anything else that we all agreed should fit into this executive?

00:34:29.87 [Michael Helbling]: I’m still a fan of the Gary Angel School of big data also being defined by not the classic BI model of how data is analyzed and worked with.

00:34:41.47 [Tim Wilson]: I’ll second that.

00:34:44.06 [Michael Helbling]: That didn’t make it into our nugget of definition from before, but it’s a nuance that I think deserves people’s attention. There’s some really great articles that Gear Angel has on this blog that I would highly recommend to anyone.

00:34:59.48 [Tim Wilson]: Which actually, I think that’s a fair point though. If you start with your business case, and you actually then you critically evaluated and Jim your example earlier where somebody says we need big data and you start asking why why why and going all Louis CK on them and they wind up what they articulate doesn’t require anything all that unique it doesn’t require necessarily a massive data set but I think when you get to a why why why that has responded with oh we do need this stuff that is a sequence of events and what’s happening in time and you have to take the amount of time in between those events and the combination in the order, then if you’ve made the business case that yes, there is something at the end of this that would allow us to make a better decision, then you’re in good shape.

00:35:51.41 [Michael Helbling]: Let’s go back in and bring something back out again and wrap it up. Jim, wrap this up for an executive.

00:35:59.94 [Jim Cain]: Well, I don’t know how I’d wrap it up for an executive, other than, and I know it was the original value proposition and now I failed, but the conversation or the discussions that we had today were really interesting to me because sometimes we agree, sometimes we don’t, but we definitely came back to a big data is not a concept that needs to be vaguely embraced by every company that can afford it. You know, it’s something that requires a specific use case. It’s something that requires a specific opportunity. Like you don’t just go, okay, we’re a hundred million dollar company, we should buy some big data. There are specific requirements. So I think the discussion with the executive is, do you know what big data is? I don’t either. So let’s figure out what you’d want to do. I actually, I was interviewed by Information Week Magazine for a piece I wrote for them like a year and change ago. And I asked the guy, I said, you know, you spend all week doing nothing but talking to people in the analytics space. Can you tell me what big data is? And he laughed and he was like, you, that’s his job. He’s like the big data, you know, editor or whatever. So an executive conversation to me is, let me help remove the fear, uncertainty and doubt from big data. The other thing is, is I think that this is going to dovetail very nicely into our conversation about data science or data scientists. Because in talking about big data, I think I finally understand the value of a data scientist. Because I’ve historically made fun of them. You guys know that. And for people listening in, I have been known to call data scientists the Y2K programmers of 2015.

00:37:40.88 [Michael Helbling]: For shame.

00:37:43.00 [Jim Cain]: For shame. But you know, we’re talking about the concept of big data. And then I think about what a data scientist does. I said it earlier. I still can’t wrap my head around a concise definition of big data, but I can now definitively tell you the value proposition of a data scientist. I don’t know if that helps.

00:38:00.35 [Tim Wilson]: So I think what we’ve made some, I think some good points a couple of times, but what we haven’t explicitly said is big data is not all hype. It’s not a, oh, big data, you know, forget about it. It’s big data is something that in some situations, And going forward, probably more of those situations, there is business value to be realized. But the number of those situations is not as high as a lot of the hype and the press and the executives feel that it is right now. So you might have a situation where big data is going to be worth the investment. Let’s sit down and figure out if that’s really the case. Have that really clear picture in your head rather than saying, I’ve just got to start doing it and I’ll figure out what I’m going to do with it later.

00:38:59.06 [Michael Helbling]: Yeah, no, I think that’s great. And that’s kind of what I’m taking away too in a certain sense, which is don’t go after big data, go after value. And as you go after value, you will start to see big data happen.

00:39:11.30 [Tim Wilson]: You may realize you’re avoiding big data. You’re avoiding big data. You’re avoiding big data. You’re chasing business value. And all of a sudden you realize you’ve been doing what other people are calling big data. Yeah. And you’re a year late into realizing that, holy crap, that’s what we’re doing.

00:39:25.07 [Michael Helbling]: But the reality is I think that’s the right perspective.

00:39:29.81 [Tim Wilson]: And that’s an ideal scenario.

00:39:32.24 [Michael Helbling]: And obviously there’s a lot more to be said on this topic. And so since there’s so much more to be said, We would love to hear from you on our Facebook page and on Twitter. If you’re an executive with C-level access to CEOs and CMOs and you want to spend a day with Tim Wilson, you have an invitation to do that. And that’s something not to look past. So thank you guys for listening, everyone. And we hope to hear from you on Facebook and on our Twitter accounts so long. And good luck with that big data out there.

00:40:09.19 [Announcer]: Thanks for listening and don’t forget to join the conversation on Facebook or Twitter. We welcome your comments and questions, facebook.com forward slash analytics hour or at analytics hour on Twitter.

Leave a Reply



This site uses Akismet to reduce spam. Learn how your comment data is processed.

Have an Idea for an Upcoming Episode?

Recent Episodes

#274: Real Talk About Synthetic Data with Winston Li

#274: Real Talk About Synthetic Data with Winston Li

https://media.blubrry.com/the_digital_analytics_power/traffic.libsyn.com/analyticshour/APH_-_Episode_274_-_Real_Talk_About_Synthetic_Data_with_Winston_Li.mp3Podcast: Download | EmbedSubscribe: RSSTweetShareShareEmail0 Shares